1. Field of Invention
Embodiments of the invention relate generally to configuration management techniques for communication networks, and more particularly, to systems and methods of connection timeout management.
2. Discussion of Related Art
A conventional communication network, such as an industrial automation network, includes at least one master device and at least one slave device. Each device communicates with other devices on the network through one or more virtual input/output (I/O) connections. Some conventional communication networks having network redundancy technologies, such as Rapid Spanning Tree Protocol (RSTP), require a non-zero amount of time to repair or reconverge from network disruptions, during which data packets are not guaranteed to reach their destinations. For example, RSTP may be configured to respond to physical link failures, such as those caused by cable or device failures, within a few milliseconds. Preferably, each device on the network should be able to tolerate or ignore network disruptions to avoid losing data, which could occur if network connections were closed prematurely.
In some conventional communication network communications protocols used in industrial automation applications, such as EtherNet/IP and Modbus TCP, each device has a timeout value associated with each I/O connection. The timeout value is a component of the network redundancy technology and may be used, at least in part, to detect and recover from lost or disrupted I/O connections. If the timeout value is too small, then the I/O connection may be terminated prematurely, the health of the connection may be incorrectly identified as bad, and/or it may take a significant amount of time for the connection to be restored. This may result in the device performing in an unexpected and/or undesirable manner during the timeout period. If the timeout value is too large, recovery from network disruptions may be delayed, negatively impacting communication speeds. Further, if the timeout value is too large, a communication problem that is not recoverable via the network redundancy technology may not be detected as quickly as can be or should be, negatively impacting the overall performance and behavior of the system. Typically, a trial-and-error technique has been used for adjusting the timeout value so that it is neither too large nor too small; however, such an approach is not optimal and is imprecise and inefficient.